Digital assistants on client devices have traditionally provided information that can be used by a user to perform a small task, such as a route to a destination location, a movie recommendation, etc. The provided information is typically generic and is not customized to a particular user. These types of systems used by digital assistants are unable to determine information that would be particularly useful to a user, and as such, the user is forced to search multiple sources, such as websites, e-mail systems, calendars, etc., to gather all information that could be used to perform a particular task. Even further, the user would need to continually search these sources to determine if anything has changed in relation to completion of a particular task. This is time consuming to the user and extremely inconvenient.